VoxRep: Enhancing 3D Spatial Understanding in 2D Vision-Language Models via Voxel Representation

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Dao, Alan, Buppodom, Norapat
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911297311866880
author Dao, Alan
Buppodom, Norapat
author_facet Dao, Alan
Buppodom, Norapat
contents Comprehending 3D environments is vital for intelligent systems in domains like robotics and autonomous navigation. Voxel grids offer a structured representation of 3D space, but extracting high-level semantic meaning remains challenging. This paper proposes a novel approach utilizing a Vision-Language Model (VLM) to extract "voxel semantics"-object identity, color, and location-from voxel data. Critically, instead of employing complex 3D networks, our method processes the voxel space by systematically slicing it along a primary axis (e.g., the Z-axis, analogous to CT scan slices). These 2D slices are then formatted and sequentially fed into the image encoder of a standard VLM. The model learns to aggregate information across slices and correlate spatial patterns with semantic concepts provided by the language component. This slice-based strategy aims to leverage the power of pre-trained 2D VLMs for efficient 3D semantic understanding directly from voxel representations.
format Preprint
id arxiv_https___arxiv_org_abs_2503_21214
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle VoxRep: Enhancing 3D Spatial Understanding in 2D Vision-Language Models via Voxel Representation
Dao, Alan
Buppodom, Norapat
Computer Vision and Pattern Recognition
Computation and Language
Comprehending 3D environments is vital for intelligent systems in domains like robotics and autonomous navigation. Voxel grids offer a structured representation of 3D space, but extracting high-level semantic meaning remains challenging. This paper proposes a novel approach utilizing a Vision-Language Model (VLM) to extract "voxel semantics"-object identity, color, and location-from voxel data. Critically, instead of employing complex 3D networks, our method processes the voxel space by systematically slicing it along a primary axis (e.g., the Z-axis, analogous to CT scan slices). These 2D slices are then formatted and sequentially fed into the image encoder of a standard VLM. The model learns to aggregate information across slices and correlate spatial patterns with semantic concepts provided by the language component. This slice-based strategy aims to leverage the power of pre-trained 2D VLMs for efficient 3D semantic understanding directly from voxel representations.
title VoxRep: Enhancing 3D Spatial Understanding in 2D Vision-Language Models via Voxel Representation
topic Computer Vision and Pattern Recognition
Computation and Language
url https://arxiv.org/abs/2503.21214